Sunteți pe pagina 1din 18

Describing System

Muhammad Ridwan Andi Purnomo, ST, MSc, PhD


Department of Industrial Engineering
Faculty of Industrial Technology
Universitas Islam Indonesia
2015

What is system
Existing definition: a group of elements that
cooperate to achieve common goals.
In term of applications, different people will give
different definition.
3 commonalities of system:
Based on cause-effect relationship
Able to be decomposed into several sub systems
Applied over a restricted application domain.

System design process


Top-down: begins with highly abstract modules and
progressively decomposes these down to an atomic
level.
Bottom-up: begins with indivisible atoms and builds
ever more abstract structures until the entire system
is defined.

System design process


Enterprise
system

system

Distributor
system

Supplier
system

Shop floor
system

Manufacturing
system

Procurement
system

system

System diagram

System properties
P1: All environmental influences on a system can be
reduced to a vector of m real variables that vary with
time, x(t) = [x1(t), . . . , xm(t)]. In general, x(t) is
called the input and the components xi(t) are input
signals.
Ex: in production shop floor, to determine number of
production, we need information on demand
Input: D. Signal of input: D1 = 150, D2 = 200, etc.

System properties
P2: All system effects can be summarized by a vector
of n real variables that vary with time, z(t) = [z1(t), . .
. , zn(t)]. In general, z(t) is called the output and the
components zi(t) are output signals.
Ex: in production shop floor, number of production (P)
is determined based on the demand.
P = output
P1 = 200, P2 = 300, P3 = 150 are the output signal.

System properties
P3: If the output signals are algebraic functions of
only the current input, the system is said to be of
zero order, since there can be no system dynamics.
Accordingly, there is a state vector y(t) = [v1](t),
yp(t)], and the system can be written as two
algebraic equations involving the input, state, and
output:
y(t) = f1(x(t))
z(t) = f2(x(t), y(t))

System properties
P4: If the input signal depends dynamically on the
output, there must also be system memory.
Ex: suppose that the system samples a signal every
t = 0, 1,2, . . . seconds and that the output z(t)
depends on input x(t- 1). It follows that there must
be two memory elements present in order to recall
x(t-1) and x(t-2) as needed. Each such implied
memory element increases the number of system
state variables by one.

System modelling diagram

Example
Example model mathematical dynamic
numerical simulation:
Consider a factory conveyor system in which boxes
arrive at the rate of one box each 10 seconds. Each
box is one of the following weights: 5, 10, or 15 kg.
However, there are twice as many 5 kg boxes and 15
kg boxes as 10 kg boxes.
How to model and simulate the system?

Example
Weight distribution of the box:
w

Pr[W=w]

0.4

10

0.2

15

0.4

Total

1.0

W is a "weight" random variable that can take on


one of the three discrete values W {5, 10,15}. The
notation Pr[W = w] is read "the probability that the
random variable W is w". The set {5,10,15} is called
the sample space of W, and is the set of all possible
weights.

Example
How many 5, 10, 15 kg boxes will come to the
system ?
The simulation procedure:
for k = l to n
r = INT(RAND()*10)+1
if r <= 4 then w(k) = 5
if 4 < r <= 6 then w(k) = 10
if r > 6 then w(k) = 15
next k

Example

Example: system with memory


Consider a factory system with two conveyors: one brings
in boxes as described in Example 1.4, but this time
arriving boxes are placed on a short conveyor that holds
exactly 3 boxes. Arriving boxes displace those on the
conveyor, which presumably just fall off! In this case, we
do not care about the individual box weight, rather we
care about the total weight on the conveyor of interest.
Thus, the input x(k) in this example can be simulated by
the output of previous example.
However, in order to determine the output g(k), the
system must remember the two previous inputs as well.
Characterize this system !

Example: system with memory


Solution:
Since the system input is a random variable, the output must
be non-deterministic as well. At the same time, after the
conveyor has been loaded, two of the three spots are known.
Thus, the second order equation will be:

Example: system with memory


The simulation:
for k=l to n
r = INT(RAND()*10)+1
if r < 4 then w(k) = 5
if 4 < r < 6 then w(k) = 10
if r > 6 then w(k) = 15
if k = l then z(k) = x(l)
if k = 2 then z(k) = x(1)+x(2)
if k > 2 then z(k) =x(k)+x(k-1)+x(k-2)
next k

THANK YOU !

S-ar putea să vă placă și